In the previously known rollover detection systems, a one-step vehicle rollover detection algorithm is used, i.e. an algorithm in which a distinction is only made between the presence of a rollover scenario and the absence of such a rollover scenario. These previously usual vehicle rollover detection algorithms include, for example, the algorithm “WinGAMR2” (windowed gyroscope algorithm measuring rollover, Version 2). The known algorithms are accordingly not in a position to distinguish rollover scenarios from almost-rollover scenarios.
A normal rollover detection is based on two paths or signals independent of one another. The first path is formed by an output signal of a rollover detection algorithm which is based on physical values from different sensors which are provided in a rollover detection module. The sensors normally include a rollover sensor, a sensor for the detection of the vertical acceleration and one or two sensors for the detection of the lateral acceleration. The relevant physical data must be subjected to a pre-processing (filtering) before they are taken account of in some mathematical and/or physical manner in the detection algorithm.
The second path can be considered as an “activation” path with basic security algorithms. An accidental rollover detection going back to a single error in the main microprocessor of the system and accordingly an accidental activation of a respective safety device should thus be prevented. Both criteria, i.e. the criteria of both paths, must therefore always be satisfied for a respective rollover detection.
A substantial requirement for a good rollover detection algorithm consists of predicting a rollover scenario within an appropriate time period (some time before the reaching of the 90° rollover angle or angle of rotation). A decision must be made between “rollover” in a rollover scenario and “no rollover” in another, normal driving situation.
In specific driving situations, it is, however, difficult to distinguish between “rollover” and “no rollover”. This is due to the fact, among other things, that in some situations the received signals are very low and develop rather slowly. Such scenarios can result in a rollover, or not.
Other situations moreover occur in which, when the signals develop very quickly and at high intensity, a rollover is expected a short time later or the vehicle falls back into its original position.
Generally, the safety of the vehicle occupants has the highest priority. If there is any doubt about the correct decision or if no clear distinction is possible between “rollover” and “no rollover”, the respective scenario should be detected as “rollover” and the safety devices should be activated.
Some safety devices are now irreversible, which means that they have to be replaced after the respective activation. This is naturally more cost intensive than the activation of reversible safety devices.